December 15, 2010
The subject of whether to upgrade datacenter backbones has been receiving extra attention since the ratification of the IEEE Std. 802.3ba standard supporting 40/100 Gigabit Ethernet back in June. It was a long time coming, but the new standard describes a single architecture capable of supporting both 40 Gigabit Ethernet (40GbE) and 100 Gigabit Ethernet (100GbE). The standard also includes a physical layer specification for communication across backplanes, copper cabling, multimode fiber and singlemode fiber. Specifically, the standard supports sending Ethernet frames at 40 and 100 gigabits per second over multiple 10 Gb/s or 25 Gb/s lanes. Prior to this declaration, the fastest standard was 10 Gigabit Ethernet, which is good-enough and cost-effective for most datacenter workloads and even some HPC workloads. Only applications that are sensitive to the latency of the underlying interconnect will need to look into lower-latency, higher bandwidth models, such as 40/100 GbE.
David Gross at DataCenterStocks.com shares his thoughts on whether upgrading makes sense: "[F]or all the talk and excitement over the new standard, the reality is that line rate is not as important a measure of network capacity as it used to be, especially with multi-gigabit port prices having more to do with transceiver reach, cabling options, and port densities than framing protocol."
In addition to which line rate you opt for, Gross asserts that there are other, perhaps equally important, decisions to be made regarding transceivers and cabling, and whether to aggregate at end-of-row or top-of-rack.
So rather than considering when 40 Gigabit Ethernet upgrades will occur, I think it's more important to monitor what's happening with average link lengths, the ratio of installed singlemode/multimode/copper ports, cross-connect densities in public data centers, the rate of transition to IPv6 peering, which can require power-hungry TCAMs within core routers, and especially whether price ratios among 850, 1310, and 1550 nanometer transceivers are growing or shrinking. So rather than wondering when 40G will achieve a 3x price ratio to 10G, it's equally important to consider whether 10G, 1550nm transceivers will ever fall below 10x the price of 10G, 850nm transceivers.
Gross also addresses the speed limitations of 40GbE:
Another factor to look at is that all the 40 Gigabit protocols showing any signs of hope within the data center run at serial rates of 10 Gigabit, whether InfiniBand or Ethernet-framed, because no one has come up with a way to run more than 10 Gigabit on a serial lane economically, even though it's been technically possible and commercially available for years on OC-768 Packet-over-SONET line cards.
Line rates are not the only consideration in choosing a datacenter backbone. Gross urges decision-makers to "look further into how other factors are now playing a large role in achieving the best cost per bit in data center networks."
Full story at DataCenterStocks.com
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?